摘要
基于四种典型舰船目标高分辨一维距离像,对特征的提取进行了探讨,分别提取了时域和变换域的特征。通过理论分析与仿真实验的方式,利用分类树、朴素贝叶斯和SVM等典型分类器对不同特征进行了识别,验证了不同分类器下使用不同特征进行目标识别的性能,通过选择分类器和基于一维距离像提取的特征,可以有效改善识别性能。
Based on the high resolution range profile(HRRP)of four typical ship targets,the feature extraction is discussed and the features of time domain and transform domain are extracted respectively.Through the way of theoretical analysis and simula⁃tion experiments,classification trees,naive bayes and SVM classifier are used to identify the different characteristics.The perfor⁃mance of using different characteristics under different classifiers is verified.By the way of selecting the classifier and extracted fea⁃tures of the targets'HRRP,the recognition performance can be effectively improved.
作者
但波
高山
戢治洪
DAN Bo;GAO Shan;JI Zhihong(Naval Aviation University,Yantai 264001)
出处
《舰船电子工程》
2020年第8期49-53,共5页
Ship Electronic Engineering
关键词
高分辨一维距离像
SVM
特征提取
目标识别
high resolution range profile
SVM
feature extraction
target recognition
